Workload-adaptive storage system with static allocation
View Patent ↗A method for adaptation of data organization by a data storage system controller is disclosed. Data organization in a data storage system is adapted according to varying workload profiles and system constraints. The data storage system also consists of a plurality of data storage units. On receipt of a request to write data, various workload and data storage system parameters are determined. On the basis of these parameters, target data storage units and appropriate data organization schemes are optimally selected. In the case where the volume of data to be written is less than a threshold value, a RAID organization scheme is preferred. In this scheme the data is written to the selected drives simultaneously. In the case where the volume of data to be written is more than the threshold value, a power managed RAID organization scheme, where all data drives need not be simultaneously powered on, is used.
1. A method for allocating storage resources in a system including a data storage system controller coupled to a data storage system, the method comprising:
determining workload parameters, wherein the workload parameters include a determined access pattern of the workload;
determining data storage system parameters, wherein the data storage system parameters include an expected I/O load per data storage unit;
using the parameters to determine if a condition is met; and
determining a configuration of the data storage system by writing the data in a power managed organizational scheme if the condition is met and by writing the data in a non-power managed organizational scheme if the condition is not met.
2. The method as recited in claim 1 , wherein determining the workload parameters comprises:
determining a volume size of the workload; and
determining an I/O block size.
3. The method as recited in claim 1 , wherein determining the workload parameters further comprises determining a target I/O rate for the workload.
4. The method as recited in claim 1 , wherein determining the workload parameters further comprises determining an expected access pattern of the workload.
5. The method as recited in claim 1 , wherein determining the data storage system parameters comprises:
determining an expected storage utilization per data storage unit for each data storage unit in the data storage system; and
determining an expected I/O bandwidth per data storage unit.
6. The method as recited in claim 1 , wherein determining the data storage system parameters further comprises determining the number of drives currently powered on per data storage unit.
7. The method as recited in claim 1 , wherein determining the data storage system parameters further comprises determining the maximum number of drives that can be powered on per data storage unit.
8. The method as recited in claim 1 , wherein determining the configuration of the data storage system:
determining target data storage unit;
determining a data organization scheme for I/O performance of volumes based on the specified parameters; and
determining a set of drives, in the target data storage unit, on which the volume is to be written.
9. The method as recited in claim 8 , wherein determining the target data storage unit is based on the lowest disk utilization.
10. The method as recited in claim 8 , wherein determining the data organization scheme comprises selecting a Redundant Array of Inexpensive Disks (RAID) organization.
11. The method as recited in claim 8 , wherein determining the data organization scheme comprises selecting a power managed Redundant Array of Inexpensive Disks (RAID) organization.
12. The method as recited in claim 8 , wherein determining the data organization scheme is based on power allocated to simultaneously power a maximum number of drives in the data storage unit.
13. The method as recited in claim 8 , wherein a determination of a set of drives is based on environmental considerations relating to electrical noise and heating.
14. The method as recited in claim 1 , further comprising allocating volumes to a location within the data storage system for continuous load balancing of the storage I/O performance of all allocated storage volumes.
15. The method of claim 1 , further comprising accepting a signal from a user input device to define at least in part an organizational scheme.
16. The method of claim 15 , wherein the accepting a signal is performed at a time of configuration of the system.
17. The method of claim 15 , wherein the accepting a signal is performed at a time of installation of the system.
18. A method for configuring I/O performance and power consumption of a data storage system on receipt of a request for writing data, the I/O performance being measured in terms of the I/O rate, the data storage system comprising a plurality of data storage units and a data storage unit controller, each of the plurality of data storage units comprising multiple disk drives, the data being written on to the disk drives, the method comprising:
determining workload parameters, wherein the workload parameters include a determined access pattern of the workload;
determining data storage system parameters, wherein the data storage system parameters include an expected I/O load per data storage unit;
identifying target data storage units for writing the data; and
if a volume of data to be written is greater than a threshold volume then writing data using a non-power managed configuration else writing data using power managed configuration.
19. The method as recited in claim 18 , wherein identifying the target data storage units comprises:
estimating possible candidate data storage units having sufficient storage capacity to allocate for writing the data; and
designating one of the possible candidate data storage units as a target data storage.
20. The method as recited in claim 18 , wherein the step of writing data using a non-power managed configuration comprises determining the number of drives for writing the data from workload parameters.